Abstract
Purpose: :
To assess the usefulness of higher–order wavefront aberration (HOA) data derived from the first corneal surface and from the whole eye for distinguishing eyes with subclinical keratoconus (KC) from normal eyes.
Methods: :
Eight clinically inconspicuous fellow eyes (paracentral inferior–superior dioptric difference [PISD] <1.4 D) of 8 patients with newly diagnosed KC in the contralateral eye were included in a prospective study. Fifteen eyes of 12 patients with newly diagnosed or mild KC without clinical signs (PISD >1.4 D) were included as positive controls and 71 healthy eyes of 40 patients served as negative controls. From axial–topographic data (Orbscan IIz, Bausch&Lomb), a Zernike decomposition of the first corneal surface was performed up to the 7th order using VOL–Pro 6.89 (Sarver and Associates). Total ocular wavefront aberrations were measured with a Hartmann–Shack sensor (Zywave, Bausch&Lomb). Both corneal and total wavefront data were normalized over a 6 mm pupil. Linear stepwise discriminant analysis was performed with input from corneal HOA data (DC), from total ocular HOA data (DO) and from both corneal and total ocular HOA data (DCO). Receiver–Operation Characteristic (ROC) curves were plotted for the output of the 3 discriminant functions. The area under the ROC curve (Az ROC) was considered as parameter for discriminative ability.
Results: :
DC, DO and DCO discriminated between normal eyes and eyes with KC with maximal sensitivity and specificity (Az ROC=1). For the discrimination of clinically inconspicuous fellow eyes (subclinical KC) from normal eyes, DCO performed best (Az ROC=1), followed by DC (Az ROC=0.99) and DO (Az ROC=0.98).
Conclusions: :
Metrics derived from discriminant analysis including HOA wavefront data of the first corneal surface and the whole eye could be useful to detect subclinical keratoconus in patients applying for refractive surgery.
Keywords: keratoconus • topography • refractive surgery: complications